Advances in Consumer Research Volume 9, 1982 Pages 389-391
CONSUMERS' INFORMATION EXPOSURE AND PATRONAGE BEHAVIOR: THREE STUDIES
Donald H. Granbois, Indiana University
The exposure phase of consumers' external information seeking is best studied in natural settings, in contrast with the usual laboratory emphasis of much consumer information processing research. The three studies discussed make some empirical and/or conceptual contributions to our further understanding of the exposure process and suggest several fruitful directions for further development and study.
It seems quite evident that the dominant emphasis in recent consumer information processing research has centered on the attention " reception b response sequence. Conceptualization and model building have been concentrated on observable, internal processes. Research studies have tended to be experimental, using messages carefully prepared or selected by the researcher and presented in controlled settings. As useful as this approach may be, it is important to remember that in the real world consumers' purchase and consumption behavior is influenced by their processing of actual messages from interpersonal and impersonal sources selected by consumers themselves. In particular, information processing must begin with exposure, a phenomenon largely outside the control of the communicator or the researcher whose objective it is to study the communication process. Whether or not exposure to a persuasive message occurs, the timing and circumstances under which exposure occurs (for example, whether an advertisement is read while casually leafing through the newspaper or actively sought out just prior to a shopping trip), and the number of times exposure occurs are all aspects of the information processing sequence which strongly influence subsequent steps. While some of the characteristics of natural exposure may be more or less simulated in controlled settings, such as the waiting room technique used in some advertising research and elsewhere, for the most part our understanding of actual exposure processes must come from studies of real consumers' actions - observed, recorded or recalled - taken in real settings.
The three papers presented in this session represent attempts to further our understanding of the exposure process. These papers are part of what may turn out to be a modest revival of interest in field studies of consumers external search processes in which the exposure phase is given some well-deserved emphasis.
Furse, Punj and Stewart have presented a large-scale and carefully designed and executed survey of pre-purchase information seeking by new automobile purchasers. Bellenger and Moschis bring a socialization perspective to the study of the patronage process, an important part of consumers' external information seeking in those frequent instances where information exposure and processing occur at least in part in the store itself. Gentry's study of the effect of credit decisions on store patronage is not as closely related to the information exposure theme, but insofar as credit does influence store visits it can be considered a study of one indirect determinant of exposure to in-store information sources.
SEARCH STRATEGIES IN NEW AUTOMOBILE PURCHASES
This relatively ambitious study used what appears to be a very carefully designed self-administered questionnaire. However, at least two methodological questions arise in considering the authors' use of the retrospective survey method
(1) How completely and accurately can respondents recall their own past behavior? Here questionnaires were received two weeks to four months after purchase (average elapsed time from purchase to receipt of questionnaire was two months), a period short enough to suggest the behavior should be reasonably fresh in respondents' minds. However, the search behavior to be recalled may have occurred some time before purchase; indeed, periods of several weeks of search and deliberation not infrequently precede new car purchases. There is an opportunity for investigation of the retrospective technique here, perhaps involving diaries or repeated surveys on information seeking among consumers actively involved in the search process for a car.
(2) Have respondents had the opportunity to adequately observe or experience all the behavior called for by the questionnaire? This question is relevant whenever more than one household member engages in information search, which is undoubtedly the case in many auto purchases as the present study suggests. The questionnaire anticipates this by asking separate questions about the respondents' behavior and that of others in the household. Validity of responses to the latter set of questions obviously rests upon the ability of the respondent (presumably the principle decision maker in each case) to accurately and fully report the behavior of others. We don't know much about this potential problem, but most people can probably think of occasions when their own behavior could not be accurately reported by another family member. The results of diary studies in which all family members keep track of information seeking during a major purchase decision could be compared with the reporting method used by the present researchers to help estimate reporting error when one respondent is used. Here again, further methodological research could help assess the degree to which the conventional retrospective reporting technique used in this and virtually all other studies o f external search is a satisfactory approach.
There are, additionally, two significant substantive aspects of the study design which deserve comment. One of these is the apparent discrepancy between the concept of "search strategies" promised in the title and abstract of the article and the reality of what was measured. "Search strategy" seems to imply a sequence of behaviors, an overall plan or heuristic which could be represented as a branching structure in which each step is contingent upon what is found at the preceding step and with one or more stopping rules present. The study's measures, however, were limited to the extent to which each information source was consulted. While this is not a serious problem in the present study, we are reminded of the lack of attention by researchers generally to the strategic or heuristic elements consumers undoubtedly develop for guiding this search process. Surely this is an important direction for future research.
The second issue is more basic and gets at the underlying theoretical basis of the study. The design seems to rest implicitly on a cost-benefit model of information search where both the value of new information and its cost guide the process The basic dependent variable, the "weights" used to indicate the relative importance of information sources, is clearly a measure of information cost. Respondents estimated the number of hours spent with each source, using a six-category scale ranging from "no time" to "over ten hours." However, no direct measures of the perceived value of new information were included. Thus it is not possible to test explicitly the notion that the wide variations in time respondents invested in pre-purchase search (ranging from an average of 6.4 hours by shoppers in Cluster 5 to 66.9 hours by shoppers in Cluster 1) is at least in part explained by differences among shoppers in the perceived value of new information, although the authors bring cost/benefit notions into their concluding discussion. One indirect measure of new information's value, satisfaction with previous car, does vary consistently in the predicted way with total hours spent in search in that the lower the satisfaction with the previous car, the more hours spent in search. The relatively small variation in satisfaction from a low mean of 4.27 to a high mean of 4.98 on a seven-point scale seems unlikely to account totally for the wide variations in hours spent, however. A second indirect indicator of the perceived value of new information--number of cars previously owned--shows no relationship with total hours invested in information search. Thus the study provides very little evidence useful in testing the Predictions of the cost/benefit model.
An additional element of the cost/benefit approach is the expectation that consumers will vary in their subjective evaluation of the cost of spending time in information search. Everyone has the same stock of time resources to allocate among all activities, yet the number of activities competing for these time allocations is likely to vary considerably from one household to another. Therefore, in very busy households, all other things equal, less time would be expected to be devoted to pre-purchase search. Again, the study design apparently doesn't include measures of time pressure and so doesn't permit testing this expected relationship.
Despite these concerns with method and scope, the study reported by Furse, Punj and Stewart appears to be an excellent contribution. Methodological refinements in future studies of this sort would seem to depend on the results of the kinds of methodological studies briefly suggested earlier. Substantive improvements center on the development of specific hypotheses aimed at explaining the intriguing very large differences in the total hours spent and the variations in the mixes of information sources consulted by consumers in the five clusters. Whether cost/benefit (this observer's choice) or some other theoretical approach is selected for testing, the finding of such large unexplained variation in search processes seems to provide strong encouragement for seeking explanation.
SOCIALIZATION APPROACHES TO PATRONAGE
Bellenger and Moschis are, understandably, unable to fully achieve in the short paper they have prepared the objective set out in their introduction. What is promised is a socialization model of retail patronage and the integration of published research with the proposed model. However, the task they describe and only incompletely pursue is important and potentially useful. While a work of at least monograph length would be needed to fully report the successful achievement of such model development and research integration, the authors should be encouraged to continue along the lines sketched out in their paper. Here are several suggestions.
(1) Explore much more fully the implications of a true socialization approach to modeling patronage behavior and provide a convincing demonstration of the unique contribution the framework provides. From the discussion in the authors' paper, one is unable to fully appreciate how the addition of cognitive development processes and the notions of socialization agents, modeling, reinforcement and social interaction can add to our understanding, ability to predict and/or influence patronage behavior.
(2) Further refine and elaborate on the model. In particular, the Behavioral Outcomes categories of General Shopping Patterns, Institutional Shopping Patterns and Store Choice seem interesting and useful. These are the prime dependent variables of interest and deserve much further development. To the extent that General Shopping Patterns encompass search strategies and heuristics, further development of this set of variables would fill a gap in the current information search literature. In fact, a real contribution could be made if interactions between patronage and search processes could be explicitlY shown.
(3) Previously-reported research studies on patronage behavior need to be more thoroughly reviewed. A more comprehensive search for such studies is needed, and more careful and complete summaries of the methods and findings of these studies is needed. Statements such as "Tate (1961) reports an inverse relationship between education and loyalty toward grocery stores" and "Similarly, Feldman and Star (1968) found racial differences in shopping behavior of Chicago shoppers" are without much value in conveying the meaning, significance and limitations of the studies quoted. Although the number of studies presented in the authors' paper is necessarily limited, it is surprising to find no evidence that they referred to literature reviews by others. It is more perplexing to find missing any reference to the few studies which have taken a socialization approach to the study of the patronage learning process of children and of "strategic populations" such as new community residents (MacNeal, 1969; Wells, 1966; Kleimenhagen and Stampfl, 1968; Bell, 1970; Atkin, 1962).
(4) Since the approach being suggested is new to most researchers in the patronage field, some rather specific examples of suggested research hypotheses and possible methods and research designs for testing them are needed.
CREDIT DECISIONS AND SHOPPING BEHAVIOR
Although Gentry doesn't emphasize the point in his discussion, the married students selected as subjects for his study constitute a "strategic population" consumers experiencing changing circumstances requiring new learning and the development of new learning patterns (Glock and Nicosia, 1963). Being granted or denied credit is surely a significant event contributing to the development of shopping patterns in recently-formed households. Thus, Gentry's study could have been an example of the kind of research the socialization model of Bellenger and Moschis can help structure. Since only a part of Gentry's study is reported in the current paper, it can't be determined whether or not other influences or further measures of developmental changes in purchasing patterns were studied.
Whether or not the study qualifies as the sort of research a socialization approach would encourage, there are some methodological aspects of his study which deserve mention. Besides the usual limitations of the college town sample to which most of us in academic research are limited, the study used a questioning procedure which resulted in responses of somewhat uncertain meaning. The weak conceptualization of the concept of "patronage" in the literature is reflected in Gentry's measure, which is operationalized as the "percentage of your purchases of goods in this store's range of merchandise." Respondents were asked to recall and estimate percentages both before and after the credit decision. Whether respondents had in mind dollars, items purchased or store visits and how accurately they categorize stores by store type remains unknown. The chances seem good that serious reporting errors, as well as differences in interpretation of the concept across respondents, have affected response. Even though the changes in patronage percentages reported tend to be in the expected direction, it could be argued that it would be remarkable if they had not been, since the patronage change measures probably better reflect attitude change than precise measures of change in actual behavior. This is not to say that the behavior did not change in the expected way; it is to say that Gentry's percentage change measures probably didn't capture these changes with much precision. Since the context in which the paper is presented is managerial, the sorts of refinements in measurement Gentry suggests deserve serious consideration in designing the larger-scale study which should follow Gentry's pilot study. Precise estimates of changes in dollar spending seem to be called for, since the goal seems to be to assess the profit implications of credit decisions.
The study could have considerably more theoretical significance had the nature of the decision process resulting in patronage of one store rather than another been considered. Does credit rejection or the unavailability of credit result in the elimination of a store from one's evoked set through some sort of conjunctive screening process? Or does it tip the balance against a store through a compensatory process? What other store choice criteria interact with credit? Clearly the store evaluation process (if indeed one occurs) should be of at least as much interest to students of consumer behavior as the brand evaluation process .
Somewhat broader implications of the credit decision which deserve study include the possible effects on the allocation of purchasing power across expenditure categories and its effects on the timing of purchases. These effects are in a broader sense, determinants of patronage behavior too and a comprehensive study of credit and patronage perhaps should include them.
Although there has been something to criticize in each of these studies, each has made some useful contribution and each represents a fruitful direction for continuing study. Papers presented at ACR conferences are often reports of work in progress or represent parts of larger projects. Audience reactions (and, hopefully, those of discussants) sometimes contribute to improvements in conceptualization and analysis as these projects progress, and audience members and proceedings readers are sometimes stimulated by the papers and others' reactions to them to modify or redirect their own research efforts.
Two concluding comments can be made in response to the three papers. Two of the papers revealed the authors to be less than completely familiar with the relevant literature, especially in the versions of the papers screened by the initial reviewers. Consumer research is a diverse and rapidly-growing field, and both the number of articles and the variety of sources in which they appear seems to increase each year. Professional associations such as ACR and the editors of our journals should be encouraged to find new ways of stimulating comprehensive literature reviews by persons willing and competent to undertake them. Researchers, in turn, should make good use of these reviews.
Secondly, the Gentry paper is typical of many others in proceedings volumes and journals in that its quality and impact could have been greatly improved had the study on which it was based been adequately funded. Consumer research, like much behavioral research, is often done on a shoestring budget by people who are capable of doing much more significant things. There is certainly no easy answer to the problem, but it seems appropriate to recognize it. Professional associations such as ACR may need to explore ways of developing mechanisms for bringing capable researchers and possible sponsors together. The academic profession may need to adjust incentive, evaluation and reward systems so that researchers in university positions are encouraged to do fewer but better funded studies.
Atkin, Kenward L., (1962), "Advertising and Store Patronage" Journal of Advertising Research, 2 (December), 18-23.
Bell, James E. Jr., (1970), "Mobiles - A Possible Segment for Retailer Cultivation," Journal of Retailing, 46 (Fall), 3-15.
Glock, Charles Y., and Nicosia, Francesco M., (1963), "Sociology and the Study of Consumers," Journal of Advertising Research, 3 (September), 21-27.
Kleimenhagen, Arno K., and Stampfl. Ronald W., (1968), "A 'Principle of Drift' for Institutional Patronage?" Journal of Retailing, 44 (Fall), 3-12.
MacNeal, James U., (1969) "An Exploratory Study of the Consumer Behavior of Children," in James U. MacNeal (ed.) Dimensions of Consumer Behavior, New York: Appleton-Century-Crofts, 255-275.
Wells, William D., (1966), "Children as Consumers," in Joseph W. Newman (ed.) On Knowing the Consumer, New York; Wiley, 138-145.